Exploring Solvable Brains: How Professor Aravinthan D.T. Samuel Integrates Physics and Neuroscience
JaKayla Harris
Dr. Aravinthan D.T. Samuel is a Professor of Physics at Harvard University, a member of the Center for Brain Science, and a leading researcher at the intersection of physics, neuroscience, and biology. His work focuses on the neural circuits that drive behavior in small organisms such as E. coli, C. elegans, and Drosophila, with the ultimate goal of extracting insights from these small brains that can further our understanding of brains in more complex organisms. Drawing on his work on bacterial behavior under Howard Berg, Dr. Samuel investigates how small organisms process sensory information and translate it into decisions and movement. His interdisciplinary lab combines techniques in microscopy, optics, machine learning, engineering, and computational modeling to link neural circuitry to behavior.
Dr. Samuel has been at Harvard for thirty-five years, where he has earned his B.A. and Ph.D., completed a postdoctoral fellowship, and become a Professor of Physics. His entire professional career has been at Harvard, where his pioneering research has been instrumental in advancing the fields of neurobiology and biophysics. In addition to his research, Dr. Samuel has developed innovative courses at the intersection of physics and biology and mentors undergraduate students, graduate students, and postdoctoral students in his lab in an effort to foster the next generation of scientists. In collaboration with his team, Dr. Samuel’s interdisciplinary, ingenious approaches to research continue to push the boundaries of what is possible in neuroscience and biophysics.
JH: What inspired you to start your work?
AS: Well, that goes back to my undergrad to work with Howard Berg. Howard Berg was the pioneer of trying to understand the organism behavior of the simplest organism you can look at. That was E coli, which is another bacteria that swims in your guts. So bacteria swimming, it does behavior right? It's moving around, swimming around, trying to do something right. It does that because it’s taking in information constantly and trying to make decisions. And bacteria, what kind of decisions does it end up making? It’s hard to say. But Howard Berg built this amazing microscope in the 60s. You can use it to follow the movements of individual bacteria in three dimensions automatically. I saw something very systematic: it runs and it tumbles – runs, tumbles, runs, tumbles – and it's like a random walk. The gradient biases the random walk, and it does that by rotating the flagella and changing the direction of the flagella. When the flagella is going counterclockwise, the bacteria swims forward, and we learn more about how when the flagella turns the way, the bacteria changes direction. Everything about the system is known in terms of inputs and outputs. Take, for instance, a nervous system. You have sensory neurons, which are chemoreceptors, and you have muscle cells, which are flagella, and you have signaling in between. With the neurons you have molecules, but a scalable system, so this is a solvable brain. Sounds like I’m enjoying my life due to a solvable brain. I’ve made progress in C. elegans and Drosophila, but I can’t say I understand anything yet. This has been worked on for fifty years, but I think we’re starting to make progress.
JH: You’re a Professor of Physics; how do you balance the interdisciplinary nature of applying physics to neuroscience?
AS: Most people in the group have a physics background: the undergraduates, graduate students, and postdocs. In this lab, we do a lot of tool building – we rarely do experiments where we use existing tools. We just build a new tool at some level, we build a tool to make a better instrument that we couldn’t make before, or we need computer science and coding to try to do data analysis faster. That kind of mindset is very typical in physics – where you want to do this experiment, but it's down the road. To get down the road, we need to do a lot of engineering first. In biology, I think people do less of that. Most of the tools are off the shelf and you try to combine them in ways for your particular biological problem. We shouldn’t have to do building – it's mostly genetics, biochemistry, and other things people know how to do. We just do it with different combinations, extensively, with instruments that have never existed before to try to make a measurement that has never been possible before.
JH: I’m curious, during undergrad and graduate school, were most of your classes in physics?
AS: Mostly, yes.
JH: Do you learn the necessary things for neuroscience on your own time?
AS: Yes. Usually, I do a lot of collaboration with those who know more biology than I do, but I am self-taught in terms of biology. One thing about building the instruments is, you ask a question that people had not asked before, or that people didn’t know they could ask before, and that’s the fun of being a biophysicist.
JH: Can you talk more about your specific research right now?
AS: Right now, with bacteria, we're trying to understand how the signaling works. The nice thing about bacteria is that ultimately, there's only one signal that's important for controlling the motor. That's a protein binding motor called CheY. This has been known for decades. If CheY binds to the motor, it makes it go clockwise. So clockwise is tumbling. If bacteria smells something it likes, it wants to go counterclockwise. What you do is lower the CheY, or CheY interaction with the motor. Howard Berg did the main experiment over 20 years ago (Berg, 2003), where he put GFPs on the motor and GFP on CheY. With CheY on the motor, they did the fluorescence attraction you could measure, and we would get a traction. You can monitor things, or you can measure the interaction. Another professor in MCB, Philippe Cluzel, did an experiment (Park et al., 2011) where you measure the CheY concentration with the bias of the cell complex – more CheY means more bias. He solved a great problem, because one thing that we now know about the system is that it's super sensitive; a small amount of CheY leads to a big behavioral change. There is neuroscience: optogenetics. You can activate or deactivate neurons with flashes of light through the activation of channels in the membrane. I said, “Well, that's all good, but what we really need is optogenetics in biochemistry where you have to activate or deactivate individual proteins,” and we invented a way to do this. We use alphafold to make a protein inactivate or deactivate with light. CheY is a signaling protein, and we can affect whether or not it binds to the motor. Now, we can research the same algorithm that applies to all the proteins inside the cell. That’s just one protein, but there are a whole bunch. What’s nice about when I started working on E. Coli, is that the brain was wired out. I knew the brain, I just had to figure out how the parts worked. It is a solvable brain. But now in neuroscience, where you are before this stage, you don’t even know what the wiring looks like. Figuring out the wiring would be connectomics. Later, during the talk, you’ll hear from Sebastian Seung about a much more challenging brain; they’re mapping a whole fly brain.
JH: How much progress has been made on mapping these brains?
AS: It depends on the organism. In C. elegans, there are only 300 neurons, and we have the whole three dimensional picture of the brain with every wire and every synapse. We didn’t just do it once, we did it across development. So that's about ten years. And we’re not the first people to do this. It was first done in the 1980s, once, and then we did it eight more times. C. elegans have 300 neurons, but a fly has 140,000 neurons. Mapping that brain took many, many more people in about ten years. For C. elegans, it was a small group effort with about ten people in ten years. The fly brain was about 100 people in ten years. Jeff Lichtman, the Dean of Science, is trying to do a large chunk of the mouse brain. We'll get to where E. coli was 30 years ago sometime in the next 30 years for bigger animals. But even now, 30 years after knowing the wiring diagram of bacteria I still don’t know how it works. All of it will come together, but in practice, we need to start with solvable problems.
.JH: Can you define what you refer to as a “solvable brain?”
AS: I originally thought that I would work on bacteria. Bacteria has been worked on for so long – decades – by Howard Berg and many other smart people in the field, so I started working on C. elegans, with brains that might be solvable and are not as big. Now, 20 years after starting work on C. elegans, I’m like, “You know, actually, it’s really hard.” We’ve made progress for sure. Then, I started working on Drosophila with 10,000 neurons. I think we’re all making progress, depending on what you define as progress. I like working with E. coli, because that’s one cell, but there are still major unresolved problems. How am I ever going to understand a system of 300 neurons when I can’t understand something with one neuron? There's so much to do, but I enjoy doing the day-to-day work that comes with trying to solve a one neuron brain.
JH: What has been your biggest challenge in your career or in your research?
AS: I like working on benchtop, small problems with a small number of people. The biggest sort of cultural challenge is that I tended to go into fields naively, instead of the normal thing – normal thing being you find someone, learn what they’re doing, become their apprentice, and train. When I started working on C. elegans, I worked in a C. elegans lab with a team of my own. Getting C. elegans people to realize what you have done is really hard. It is the same for Drosophila. What’s weird is that now I’m doing it again, with bacteria. Even though I did work on bacteria as a graduate student, it's been some time, so everyone has forgotten who I am. Now, I have to get back into my old field. So, the challenge and the fun is trying to get people in new fields to notice you as a newcomer. Another nice thing is that people will leave you alone a little for a little bit; I’m not a very social person.
JH: Where do you see this field going in the future?
AS: Hopefully, we'll get the understanding of animals with real brains – multineural brains. Hopefully, we’ll someday get to the point where we understand the E. coli brain. At hand, I’m still waiting to solve the E. coli brain, so I think it will take longer. A holistic understanding, all the way from input to output, seems unimaginable in anything larger than a fly brain.
JH: What is your opinion on this research being done on the human brain?
AS: That’s pointless. For sure, understanding small brains will give you insight into how big brains work. It’s fair to say that the single organism that has taught us more about human biology than any other organism is E. coli. Everything we know about genetics and microbiology – gene regulation, transcriptional switches, etc. – is all E. coli. Signal processing, information processing – that's where we have a consensus. So, if you understand E. coli really well, you might have a chance of understanding an organism that is a little bigger. People say, “Oh you should try a mouse because it’s more similar since it’s mammalian,” but it’s also unsolvable. I’m sure there’s been a lot of mileage that came from mice, but the total aggregate of understanding is not even close to what E. coli has told us anymore.
JH: Is there anything else you would like to talk about?
AS: When I started here at 18 years old, I was an undergraduate researcher in a lab. And here I still am, 35 years later. Now, I’ve been here so long I can’t imagine leaving. I’ve been here continuously for 35 years.
About the Author
Jakayla Harris (‘27) is a sophomore at Harvard College concentrating in physics.
Additional Links
Dr. Samuels Website: https://samuel.physics.harvard.edu/
References
Dr. Samuel has been at Harvard for thirty-five years, where he has earned his B.A. and Ph.D., completed a postdoctoral fellowship, and become a Professor of Physics. His entire professional career has been at Harvard, where his pioneering research has been instrumental in advancing the fields of neurobiology and biophysics. In addition to his research, Dr. Samuel has developed innovative courses at the intersection of physics and biology and mentors undergraduate students, graduate students, and postdoctoral students in his lab in an effort to foster the next generation of scientists. In collaboration with his team, Dr. Samuel’s interdisciplinary, ingenious approaches to research continue to push the boundaries of what is possible in neuroscience and biophysics.
JH: What inspired you to start your work?
AS: Well, that goes back to my undergrad to work with Howard Berg. Howard Berg was the pioneer of trying to understand the organism behavior of the simplest organism you can look at. That was E coli, which is another bacteria that swims in your guts. So bacteria swimming, it does behavior right? It's moving around, swimming around, trying to do something right. It does that because it’s taking in information constantly and trying to make decisions. And bacteria, what kind of decisions does it end up making? It’s hard to say. But Howard Berg built this amazing microscope in the 60s. You can use it to follow the movements of individual bacteria in three dimensions automatically. I saw something very systematic: it runs and it tumbles – runs, tumbles, runs, tumbles – and it's like a random walk. The gradient biases the random walk, and it does that by rotating the flagella and changing the direction of the flagella. When the flagella is going counterclockwise, the bacteria swims forward, and we learn more about how when the flagella turns the way, the bacteria changes direction. Everything about the system is known in terms of inputs and outputs. Take, for instance, a nervous system. You have sensory neurons, which are chemoreceptors, and you have muscle cells, which are flagella, and you have signaling in between. With the neurons you have molecules, but a scalable system, so this is a solvable brain. Sounds like I’m enjoying my life due to a solvable brain. I’ve made progress in C. elegans and Drosophila, but I can’t say I understand anything yet. This has been worked on for fifty years, but I think we’re starting to make progress.
JH: You’re a Professor of Physics; how do you balance the interdisciplinary nature of applying physics to neuroscience?
AS: Most people in the group have a physics background: the undergraduates, graduate students, and postdocs. In this lab, we do a lot of tool building – we rarely do experiments where we use existing tools. We just build a new tool at some level, we build a tool to make a better instrument that we couldn’t make before, or we need computer science and coding to try to do data analysis faster. That kind of mindset is very typical in physics – where you want to do this experiment, but it's down the road. To get down the road, we need to do a lot of engineering first. In biology, I think people do less of that. Most of the tools are off the shelf and you try to combine them in ways for your particular biological problem. We shouldn’t have to do building – it's mostly genetics, biochemistry, and other things people know how to do. We just do it with different combinations, extensively, with instruments that have never existed before to try to make a measurement that has never been possible before.
JH: I’m curious, during undergrad and graduate school, were most of your classes in physics?
AS: Mostly, yes.
JH: Do you learn the necessary things for neuroscience on your own time?
AS: Yes. Usually, I do a lot of collaboration with those who know more biology than I do, but I am self-taught in terms of biology. One thing about building the instruments is, you ask a question that people had not asked before, or that people didn’t know they could ask before, and that’s the fun of being a biophysicist.
JH: Can you talk more about your specific research right now?
AS: Right now, with bacteria, we're trying to understand how the signaling works. The nice thing about bacteria is that ultimately, there's only one signal that's important for controlling the motor. That's a protein binding motor called CheY. This has been known for decades. If CheY binds to the motor, it makes it go clockwise. So clockwise is tumbling. If bacteria smells something it likes, it wants to go counterclockwise. What you do is lower the CheY, or CheY interaction with the motor. Howard Berg did the main experiment over 20 years ago (Berg, 2003), where he put GFPs on the motor and GFP on CheY. With CheY on the motor, they did the fluorescence attraction you could measure, and we would get a traction. You can monitor things, or you can measure the interaction. Another professor in MCB, Philippe Cluzel, did an experiment (Park et al., 2011) where you measure the CheY concentration with the bias of the cell complex – more CheY means more bias. He solved a great problem, because one thing that we now know about the system is that it's super sensitive; a small amount of CheY leads to a big behavioral change. There is neuroscience: optogenetics. You can activate or deactivate neurons with flashes of light through the activation of channels in the membrane. I said, “Well, that's all good, but what we really need is optogenetics in biochemistry where you have to activate or deactivate individual proteins,” and we invented a way to do this. We use alphafold to make a protein inactivate or deactivate with light. CheY is a signaling protein, and we can affect whether or not it binds to the motor. Now, we can research the same algorithm that applies to all the proteins inside the cell. That’s just one protein, but there are a whole bunch. What’s nice about when I started working on E. Coli, is that the brain was wired out. I knew the brain, I just had to figure out how the parts worked. It is a solvable brain. But now in neuroscience, where you are before this stage, you don’t even know what the wiring looks like. Figuring out the wiring would be connectomics. Later, during the talk, you’ll hear from Sebastian Seung about a much more challenging brain; they’re mapping a whole fly brain.
JH: How much progress has been made on mapping these brains?
AS: It depends on the organism. In C. elegans, there are only 300 neurons, and we have the whole three dimensional picture of the brain with every wire and every synapse. We didn’t just do it once, we did it across development. So that's about ten years. And we’re not the first people to do this. It was first done in the 1980s, once, and then we did it eight more times. C. elegans have 300 neurons, but a fly has 140,000 neurons. Mapping that brain took many, many more people in about ten years. For C. elegans, it was a small group effort with about ten people in ten years. The fly brain was about 100 people in ten years. Jeff Lichtman, the Dean of Science, is trying to do a large chunk of the mouse brain. We'll get to where E. coli was 30 years ago sometime in the next 30 years for bigger animals. But even now, 30 years after knowing the wiring diagram of bacteria I still don’t know how it works. All of it will come together, but in practice, we need to start with solvable problems.
.JH: Can you define what you refer to as a “solvable brain?”
AS: I originally thought that I would work on bacteria. Bacteria has been worked on for so long – decades – by Howard Berg and many other smart people in the field, so I started working on C. elegans, with brains that might be solvable and are not as big. Now, 20 years after starting work on C. elegans, I’m like, “You know, actually, it’s really hard.” We’ve made progress for sure. Then, I started working on Drosophila with 10,000 neurons. I think we’re all making progress, depending on what you define as progress. I like working with E. coli, because that’s one cell, but there are still major unresolved problems. How am I ever going to understand a system of 300 neurons when I can’t understand something with one neuron? There's so much to do, but I enjoy doing the day-to-day work that comes with trying to solve a one neuron brain.
JH: What has been your biggest challenge in your career or in your research?
AS: I like working on benchtop, small problems with a small number of people. The biggest sort of cultural challenge is that I tended to go into fields naively, instead of the normal thing – normal thing being you find someone, learn what they’re doing, become their apprentice, and train. When I started working on C. elegans, I worked in a C. elegans lab with a team of my own. Getting C. elegans people to realize what you have done is really hard. It is the same for Drosophila. What’s weird is that now I’m doing it again, with bacteria. Even though I did work on bacteria as a graduate student, it's been some time, so everyone has forgotten who I am. Now, I have to get back into my old field. So, the challenge and the fun is trying to get people in new fields to notice you as a newcomer. Another nice thing is that people will leave you alone a little for a little bit; I’m not a very social person.
JH: Where do you see this field going in the future?
AS: Hopefully, we'll get the understanding of animals with real brains – multineural brains. Hopefully, we’ll someday get to the point where we understand the E. coli brain. At hand, I’m still waiting to solve the E. coli brain, so I think it will take longer. A holistic understanding, all the way from input to output, seems unimaginable in anything larger than a fly brain.
JH: What is your opinion on this research being done on the human brain?
AS: That’s pointless. For sure, understanding small brains will give you insight into how big brains work. It’s fair to say that the single organism that has taught us more about human biology than any other organism is E. coli. Everything we know about genetics and microbiology – gene regulation, transcriptional switches, etc. – is all E. coli. Signal processing, information processing – that's where we have a consensus. So, if you understand E. coli really well, you might have a chance of understanding an organism that is a little bigger. People say, “Oh you should try a mouse because it’s more similar since it’s mammalian,” but it’s also unsolvable. I’m sure there’s been a lot of mileage that came from mice, but the total aggregate of understanding is not even close to what E. coli has told us anymore.
JH: Is there anything else you would like to talk about?
AS: When I started here at 18 years old, I was an undergraduate researcher in a lab. And here I still am, 35 years later. Now, I’ve been here so long I can’t imagine leaving. I’ve been here continuously for 35 years.
About the Author
Jakayla Harris (‘27) is a sophomore at Harvard College concentrating in physics.
Additional Links
Dr. Samuels Website: https://samuel.physics.harvard.edu/
References
- Berg, H. C. (2003). The rotary motor of bacterial flagella. Annual Review of Biochemistry, 72(1), 19–54. https://doi.org/10.1146/annurev.biochem.72.121801.161737
- Park, H., Oikonomou, P., Guet, C. C., & Cluzel, P. (2011). Noise underlies switching behavior of the bacterial flagellum. Biophysical Journal, 101(10), 2336–2340. https://doi.org/10.1016/j.bpj.2011.09.040